广告
加载中

宇树科技IPO倒计时 9000万撑起的千亿帝国 还差一个“大脑”

张申宇 2026-05-27 08:48
张申宇 2026/05/27 08:48

邦小白快读

EN
全文速览

这篇文章核心介绍了即将上会IPO的宇树科技的发展现状、核心优势与未来挑战,核心干货如下:

1. 核心基础信息:宇树科技成立9年,是A股最有分量的人形机器人预备IPO企业,拟募资42亿元,发行后对应市值420亿元,上市后有望冲击千亿;2025年营收突破17亿元,同比增长超335%,扣非净利润超6亿元,人形机器人出货量超5500台居全球第一。

2. 现有核心优势:走全栈自研+小脑(运动控制)优先的路线,四足机器人全球市占近60%,人形机器人全球市占32.4%,靠高性价比快速占领市场,已经实现盈利爆发。

3. 现存核心短板:年研发投入仅约9000万元,自研通用具身大模型尚未规模化应用,本次IPO募资超20亿元将投入大脑技术研发,技术突破成败决定未来发展。

本文披露了人形机器人头部品牌宇树科技的发展路径与赛道趋势,对消费机器人领域品牌商有诸多参考干货,具体如下:

1. 品牌营销可借鉴:宇树依靠产品性能优势登上春晚舞台,靠刷屏事件快速打出品牌势能,抢占用户心智,用极低的营销成本获得了广泛的大众认知,适合新消费科技品牌参考。

2. 战略与定价优势:宇树在行业大模型技术尚未成熟的阶段,选择优先发展可落地的运动控制技术,通过全栈自研替代外部采购,压缩成本打造性价比壁垒,快速占领市场,匹配当前市场对高性能低价机器人的需求。

3. 行业趋势:当前人形机器人已经进入量产元年,需求爆发式增长,下一步核心竞争点将转向认知AI能力,品牌需要提前布局相关技术储备。

本文披露了人形机器人赛道的最新发展情况,对布局机器人相关赛道的卖家有诸多干货参考,内容如下:

1. 市场机会判断:当前全球人形机器人已经进入量产元年,市场需求呈现数倍爆发式增长,2025年仅宇树一家人形机器人出货就超5500台,全球每三台就有一台出自宇树,四足和人形双赛道都有稳定的市场需求,赛道增长空间广阔。

2. 可借鉴的商业模式:宇树走全栈自研控制成本,优先发展成熟落地的运动控制技术,靠高性价比快速抢占市占,先实现盈利再攻坚前沿技术,这条路已经被验证成功,适合布局赛道的卖家参考。

3. 风险提示:当前赛道核心瓶颈是具身大模型技术,没有核心AI能力的玩家后续很容易被淘汰,赛道竞争会越来越激烈,需要提前布局核心技术。

本文披露了宇树科技的生产研发模式,对机器人及相关制造工厂有诸多启示,干货内容如下:

1. 产品生产设计方向:当前市场对机器人的运动控制能力、稳定性要求已经很高,用户愿意为能完成高难度动作的高性能产品买单;宇树通过全栈自研所有核心部件,沉淀工程经验并掌控供应链,有效压缩生产成本,这种模式值得制造工厂借鉴。

2. 商业机会:当前人形机器人已经进入量产爆发期,出货量每年数倍增长,核心零部件需求大幅提升,给上游零部件生产工厂带来了大量增量订单机会。

3. 转型启示:传统工厂转型机器人领域,不需要一开始就投入巨资砸尚未成熟的通用大模型,先从成熟的硬件工程化能力切入,做出性价比产品占领市场,再逐步攻坚前沿技术,路径更务实。

本文分析了当前具身智能人形机器人赛道的发展现状与核心痛点,对服务机器人行业的服务商有诸多参考干货,内容如下:

1. 行业发展趋势:全球人形机器人已经进入量产元年,出货量快速增长,目前头部厂商已经基本解决了运动控制和本体硬件的问题,接下来整个行业的核心发展方向是具身大模型的研发与落地,行业即将迎来从“表演级”到“实用级”的质变。

2. 核心客户痛点:当前整个机器人行业的最大技术瓶颈就是可落地的通用具身大模型,也就是机器人的认知“大脑”,现有产品大多还是硬件开发平台,无法完成复杂场景的自主决策,严重制约了商业化落地。

3. 市场机会:接下来大量机器人厂商会投入巨资研发具身大模型和行业解决方案,服务商可以围绕大模型训练、算力服务、场景落地解决方案等方向布局,抢占赛道增量。

本文披露了人形机器人赛道头部企业的IPO动向,对服务硬科技产业的平台商有诸多参考干货,内容如下:

1. 产业需求:当前人形机器人赛道头部企业已经进入资本化阶段,企业需要大量资金支持前沿核心技术研发,宇树本次IPO拟募资42亿元,其中20亿元直接用于具身大模型研发,说明产业十分需要资本平台支持技术攻坚。

2. 可借鉴的最新做法:上交所科创板针对优质硬科技企业推出预先审阅机制,大幅提升了审核效率,宇树作为该机制下第二单申报企业,享受到了更快的审核进度,有利于优质企业快速融资发展,这个做法值得各类产业平台参考。

3. 风向规避:平台在引入机器人赛道企业时,需要关注企业研发投入与技术储备的匹配性,要提示核心技术短板的风险,引导企业聚焦核心技术研发,避免行业出现泡沫。

本文披露了人形机器人赛道头部企业宇树科技的最新发展动向,为具身智能产业研究提供了典型样本,核心干货如下:

1. 产业新动向:当前全球人形机器人已经进入量产元年,中国厂商已经在出货量上占据全球领先位置,宇树人形机器人出货量全球第一,四足机器人市占接近60%,产业已经从研发试验阶段进入量产盈利阶段,技术路线分化为认知“大脑”和运动“小脑”两个方向。

2. 创新商业模式研究:宇树走出了差异化发展路径,通过全栈自研降本,先做成熟的小脑技术抢占市场,实现盈利爆发后再募资攻坚大模型技术,和其他同行形成差异化,这个新模式为产业发展提供了新的研究样本。

3. 政策启示:监管层已经通过预先审阅机制提升优质硬科技企业的IPO效率,支持机器人产业发展,后续仍需要出台针对性政策支持具身大模型研发,帮助产业突破核心技术瓶颈。

返回默认

声明:快读内容全程由AI生成,请注意甄别信息。如您发现问题,请发送邮件至 run@ebrun.com 。

我是 品牌商 卖家 工厂 服务商 平台商 研究者 帮我再读一遍。

Quick Summary

This article provides a core overview of Unitree Robotics, a Chinese humanoid robot company soon to be listed via IPO, covering its current development, core strengths and future challenges. Key takeaways are as follows:

1. Core background: Founded nine years ago, Unitree is the most prominent pre-IPO humanoid robot candidate for China's A-share market. It plans to raise RMB 4.2 billion via the offering, implying a post-issue market cap of RMB 42 billion, with potential to reach RMB 100 billion post-listing. In 2025, the company posted revenue exceeding RMB 1.7 billion, up over 335% year-on-year, with non-GAAP net profit above RMB 600 million. Its humanoid robot shipments surpassed 5,500 units, ranking first globally.

2. Core competitive strengths: Unitree follows a full-stack self-development strategy prioritizing "cerebellum" (motion control) technology. It holds nearly 60% global market share in quadruped robots and 32.4% in humanoid robots, capturing market share rapidly via high cost-performance, and has already achieved explosive profit growth.

3. Key existing weaknesses: Unitree's annual R&D investment is only around RMB 900 million. Its self-developed general embodied large model has not yet reached large-scale commercial application. Over RMB 2 billion of the IPO proceeds will be allocated to "brain" (cognitive AI) technology R&D, and the success of this technological breakthrough will determine its future development.

This article outlines Unitree Robotics' growth path and sector trends, offering actionable insights for brand owners in the consumer robotics space. Key takeaways are as follows:

1. Brand marketing lessons: Unitree leveraged its product performance advantage to secure a spot at China's annual Spring Festival Gala, building strong brand momentum and capturing consumer mindshare through viral exposure. It achieved widespread public recognition with extremely low marketing costs, a model worth emulating for new consumer technology brands.

2. Strategic and pricing advantages: At a stage when industry large model technology is still immature, Unitree chose to prioritize commercially viable motion control technology first. It replaced external component procurement with full-stack in-house development to cut costs and build a cost-performance moat, allowing it to capture market quickly while meeting current market demand for high-performance, low-priced robots.

3. Sector trend outlook: The humanoid robot industry has now entered its first year of mass production, with demand growing explosively. The next core competitive battleground will shift to cognitive AI capabilities, and brands need to build relevant technology reserves in advance.

This article shares the latest developments in the humanoid robot track, offering valuable insights for sellers active in robotics-related segments. Key takeaways are as follows:

1. Market opportunity assessment: The global humanoid robot industry has now entered its first year of mass production, with demand growing exponentially. Unitree alone shipped over 5,500 humanoid robots in 2025, accounting for one out of every three humanoid robots sold globally. Both the quadruped and humanoid robot segments enjoy stable market demand, with broad room for future growth.

2. Proven business model lessons: Unitree cuts costs via full-stack self-development, prioritizes mature, commercially viable motion control technology, captures market share rapidly via high cost-performance, and achieves profitability before investing in cutting-edge technology. This path has been proven successful and is well worth referencing for sellers entering the track.

3. Risk warning: The current core bottleneck of the track is embodied large model technology. Players without core AI capabilities are very likely to be phased out as competition intensifies, so players need to布局 core technology in advance.

This article analyzes Unitree Robotics' R&D and production model, offering key insights for robotics and related manufacturing factories. Key takeaways are as follows:

1. Product design and production guidance: The market already has high requirements for robots' motion control capability and stability, and consumers are willing to pay for high-performance products that can complete complex movements. Unitree develops all core components in-house via full-stack self-research, accumulates engineering experience, controls its supply chain, and effectively reduces production costs – a model worthy of reference for manufacturing factories.

2. Commercial opportunities: The humanoid robot industry is now in a mass production boom, with shipments growing exponentially every year. Demand for core components has risen sharply, bringing a large volume of incremental order opportunities for upstream component manufacturing factories.

3. Transformation insights: For traditional factories transforming into the robotics sector, it is not necessary to invest heavily in immature general large models from the start. A more pragmatic path is to enter the market via mature hardware engineering capabilities, capture market share with cost-effective products, and gradually pursue breakthroughs in cutting-edge technology.

This article analyzes the current development status and core pain points of the embodied intelligent humanoid robot track, offering key takeaways for service providers in the service robotics industry. Key insights are as follows:

1. Industry development trends: The global humanoid robot industry has entered its first year of mass production, with shipments growing rapidly. Leading players have already largely solved challenges around motion control and本体 hardware. The next core development direction for the entire industry is the R&D and commercialization of embodied large models, and the sector is about to undergo a qualitative shift from "demonstration-grade" to "utility-grade" products.

2. Core customer pain points: The biggest technological bottleneck facing the entire robotics industry today is the development of deployable general embodied large models – the cognitive "brain" of robots. Most existing products remain hardware development platforms, incapable of autonomous decision-making in complex scenarios, which severely limits commercial落地.

3. Market opportunities: A large number of robot manufacturers will invest heavily in embodied large model development and industry-specific solutions in the coming period. Service providers can布局 around large model training, computing power services, scenario-based落地 solutions and other related directions to capture incremental growth in the track.

This article covers the upcoming IPO of a leading humanoid robot company, offering insights for platform operators serving the hard tech sector. Key takeaways are as follows:

1. Industry demand: Leading players in the humanoid robot track have now entered the capitalization stage, and require large amounts of capital to support cutting-edge core technology R&D. Unitree plans to raise RMB 4.2 billion in its IPO, of which RMB 2 billion will be directly allocated to embodied large model R&D, demonstrating that the industry strongly needs capital platform support for technological breakthroughs.

2. Referenceable best practices: The Shanghai Stock Exchange Science and Technology Innovation Board (STAR Market) has launched a pre-review mechanism for high-quality hard tech enterprises, which greatly improves review efficiency. As the second enterprise to file under this mechanism, Unitree has benefited from a faster review process that allows high-quality companies to secure financing and grow more quickly. This approach is worthy of reference for all types of industrial platforms.

3. Risk guidance: When onboarding robot track enterprises, platforms should assess the alignment between enterprises' R&D investment and technology reserves, disclose risks related to core technology weaknesses, guide enterprises to focus on core technology R&D, and prevent sector bubbles.

This article presents the latest developments of Unitree Robotics, a leading player in the humanoid robot sector, offering a representative case study for embodied intelligence industry research. Key insights are as follows:

1. New industry trends: The global humanoid robot industry has entered its first year of mass production, and Chinese manufacturers already hold a global leading position in shipment volume. Unitree ranks first globally in humanoid robot shipments and holds nearly 60% global market share in quadruped robots. The industry has transitioned from the R&D and testing phase to the mass production and profitability phase, and technology routes have diverged into two directions: cognitive "brain" and motion "cerebellum".

2. Innovative business model research: Unitree has carved out a differentiated development path: it cuts costs via full-stack self-development, first captures market share with the mature cerebellum technology, achieves profitable growth, and then raises capital to pursue large model technology breakthroughs. This new differentiated model provides a fresh research sample for industry development.

3. Policy implications: Regulators have already improved IPO efficiency for high-quality hard tech enterprises via the pre-review mechanism to support the robotics sector's growth. Going forward, targeted policies are still needed to support embodied large model R&D and help the industry break through core technology bottlenecks.

Disclaimer: The "Quick Summary" content is entirely generated by AI. Please exercise discretion when interpreting the information. For issues or corrections, please email run@ebrun.com .

I am a Brand Seller Factory Service Provider Marketplace Seller Researcher Read it again.

将于6月1日上会审议的宇树科技,已经走到了从“明星创业公司”到“上市公司”的历史分水岭。

5月25日,上交所官网一则公告,将沉寂数月的具身智能赛道再次推向聚光灯下——宇树科技股份有限公司科创板IPO将于6月1日上会审议,公司拟公开发行新股不低于4044.64万股,计划募集资金42.02亿元,按发行底价测算,其对应市值为420亿元,上市后有望冲击千亿。宇树科技也成为继优必选(港股已上市)、越疆之后,A股最具分量的“人形机器人第一股”预备役选手。

从四足机器狗登顶春晚舞台,到人形机器人以每年数倍量级爆发式增长,短短九年间,宇树科技已然构建起一座外人难以复制的产品帝国。但翻看其长达数百页的招股说明书,一组数字让不少人倒吸了一口凉气——2024年度,宇树科技研发费用仅9046.63万元;2025年前三季度,研发费用9020.94万元。

9000万元甚至不及国内一些AI初创企业一年的算力开销。而就在不久前的公开演讲中,宇树科技创始人王兴兴对外透露,公司2025年营收已突破17.08亿元,同比增长超过335%,扣非后净利润逾6亿元,人形机器人出货量超5500台,位居全球第一。

营收利润暴增、出货量全球登顶,研发投入却看上去“不太相称”——这是外界最大的疑惑,也是贯穿宇树本次IPO的核心悬念:一家年研发费用不足亿元的企业,凭什么在激战正酣的机器人赛道中守住护城河,并敢向特斯拉叫板?而在6月1日上会之后,手握42亿元募资弹药的王兴兴,能否补上那最后一块“大脑”短板?

小脑+本体的优势,让宇树赢下上半场

2025年,全球人形机器人赛道迎来了“量产元年”的历史节点。这其中,宇树科技以超5500台的出货量居全球首位,全球占比达32.4%。换言之,全球每卖出三台人形机器人,就有一台来自宇树。招股书同时显示,报告期内其四足机器人销量合计已超3万台,全球市占率接近60%。

但在其不到一个月前公布的招股书中,9000万的研发成本却让很多业内外人员对其未来,以及当下的繁荣是否“虚假”提出了质疑。外界甚至将其解读为“从技术驱动向市场驱动过早转向”的风险信号。

宇树科技研发费用率为何远低于同行业平均水平?宇树科技主要依靠两个核心:“全栈自研”压缩成本 + “小脑优先”抢占市场。

“全栈自研”方面,即从机器人本体、核心算法到关节模组、激光雷达等关键部件全链条自主研发制造。这种模式的短期研发成本看似“不够高”,是因为大量对外采购的“黑盒”模块被自研替代,研发投入沉淀为工程Know-How和供应链深度掌控,而非高额的外部专利费用。

小脑方面,在具身智能“大脑”(认知推理)与“小脑”(运动控制)两条技术路线中,宇树在前期几乎将全部研发资源押注于后者。这并非战略失误,而是基于行业现实作出的务实选择:在全球具身大模型技术均处研发测试阶段的背景下,抢先打造一块令对手望尘莫及的“肌肉记忆”,成为机器人进入复杂物理环境的唯一入场券。

事实证明,宇树的选择踩准了节奏。H1以超5米/秒奔跑速度刷新全尺寸人形机器人世界纪录,并完成了全球首例电驱原地后空翻等高难度动作。在2026年春晚舞台上,宇树的多款主力机型甚至在高强度编排中展现出高度的协同与平衡,远超行业平均水平。

与此同时,宇树与云深处、乐聚等其他拟上市同行形成了鲜明的“产品组合”差异。在市场关注度最高的四足机器狗与人形机器人双赛道中,宇树是为数不多的同时占据“双冠军”的企业。据刚刚过去的2026年5月多家拟上市机器人公司招股书横向对比来看,宇树在营收规模、净利润、人形机器人出货量等三个维度上均大幅领先云深处和乐聚等竞争者。

可以说,宇树能在研发投入相对有限的情况下率先实现营收和盈利的爆发,背后是“技术选择”与“商业节奏”的高度匹配——用极致的工程化能力构建出别人难以复制的“性价比壁垒”,再以高性价比产品迅速占领用户心智并提升市占率。这条路径在消费电子与新能源车行业已被反复验证,如今宇树正在机器人赛道复刻类似的剧本。

“大脑”或将成为其“木桶短板”

然而,在本体+小脑成功的背后,或许已经暗藏隐患。当全球机构投资者翻开宇树科技的招股说明书时,最揪心的并不是那张亮眼的财报,而是一条低调却极为关键的警告——招股书中明确提示,“全球具身大模型技术尚处研发测试阶段,公司报告期内尚未将自研的通用具身大模型规模化应用于机器人产品。若‘大脑’技术未能取得重要进展,将使得通用机器人的大规模应用进程存在不确定性。”

王兴兴对此并不讳言。在不久前的公开访谈中,他曾公开表示,“具身智能真正的‘GPT时刻’目前还差一点火候。行业里有些人比较乐观,预估18个月就能实现,我可能稍微悲观一点,觉得至少需要两到3年的时间,但过程肯定也会非常快。”

但这种坦诚并不能消弭资本市场的担忧。因为一个无法回避的事实是:迄今为止,宇树交付给客户的机器人,更多仍是高性能的“硬件开发平台”,可跑、可跳、可后空翻,但尚不能完全自主理解复杂指令、进行多步骤任务规划并应对未知环境的变化。

这种能力缺失,恰是具身智能产业界当前最大的技术瓶颈。早在今年2月,高盛就在发布的最新研报中指出,尽管中国人形机器人硬件工程取得了显著进步,但“真实AI能力仍待检验”。高盛同时提出,长期突破将取决于AI“世界模型”的技术进展,它将直接决定2035年138万台人形机器人目标的可达性。

从产品落地角度看,这一短板已经对宇树的商业化推展形成了潜在制约。招股书中明确提到,公司来自四足机器人的收入占比正在逐步下降,而人形机器人收入占比从2023年的1.88%快速攀升至2025年前三季度的51.53%——这意味着,人形机器人正成为宇树未来最重要的增长引擎。但人形机器人要实现从“表演级”到“实用级”的质变,最终走向非标工业装配乃至家庭服务场景,必然依赖于能够真正理解环境并自主决策的“大脑”。

这正是为什么,在宇树此次IPO的42.02亿元募资计划中,最重头的一个去向就是“智能机器人模型研发项目”,投资总额高达20.22亿元,位居四项募投项目之首。

将于6月1日上会审议的宇树科技,已经走到了从“明星创业公司”到“上市公司”的历史分水岭。

值得关注的是,这并非一次普通的科创板IPO。据上海证券报此前的报道,宇树科技是科创板预先审阅机制下的第二单申报企业,此前已完成两轮问询答复。这一身份不仅体现了监管层对机器人产业链重要性的高度认可,也意味着宇树在IPO审核效率上具备显著优势,有望成为A股最快上市的机器人公司之一。

过去,宇树的竞争壁垒建立在两个传统优势之上:一是以“全栈自研”和“极致成本”构建起的行业价格防线;二是以春晚刷屏等品牌势能抢占的用户心智。这两者共同构成了业内常说的“性价比壁垒”。这套打法在跑通消费级市场、实现产品快速铺量方面被证明是有效的,其财务成果也体现在了招股书的高速增长数据中。

然而,二级市场的法则与风险投资的逻辑截然不同。当股价走势取代营销声量成为最大的现实压力,宇树被迫需要回答的,绝不只是“一年能卖多少台”,而是“三年后还能不能继续卖这么多”。

换句话说,资本市场关心的不是宇树过去的成功,而是它在未来十年持续成功的可能性。而对于宇树科技而言,42亿元的巨额募资,既是护身符,也是赌注。如果这笔资金能够帮助宇树攻克具身大模型的技术高峰、扩展行业解决方案深度并在全球竞争中构建真正的“生态级壁垒”,那么今天对9000万研发费用的所有质疑都将不值一提。

反之,如果宇树在上市后未能证明其拥有从“小脑”迈向“大脑”的跨越能力,或者在市场格局剧变中丧失了成本护城河的领先优势,那么今天招股书中那些光鲜的数据,就可能从“优势”异化为“负担”,毕竟如今“野蛮生长”的人形机器人行业是不会给后来者太多等待的时间。

注:文/张申宇,文章来源:钛媒体(公众号ID:taimeiti),本文为作者独立观点,不代表亿邦动力立场。

文章来源:钛媒体

广告
微信
朋友圈

这么好看,分享一下?

朋友圈 分享

APP内打开

+1
+1
微信好友 朋友圈 新浪微博 QQ空间
关闭
收藏成功
发送
/140 0